THE MARGIN PLAYBOOK
You quote project work. Some of it earns less than you priced.
This is how to find out what they actually earned.
Four minutes to read. Four to five days to run properly, and the work is rather the point.
START HERE
Check one finished job
Compare the gross margin you planned with the gross margin the job delivered. Use the same delivery-cost basis for both figures. We do not apply an industry benchmark.
The calculation runs in your browser. The figures are not submitted or tracked.
YOUR RESULT
Use a completed job, not an industry average.
If you cannot find the actual delivery cost yet, that is the first useful finding. Move 1 shows where to pull it from.
See Move 1THE PLAYBOOK
Five moves, and none of them need us
Every move below runs in a spreadsheet. The worked example is an invented commercial fit-out. Its figures come from one tested model, so they stay consistent throughout.
Run properly across ten jobs, this is four to five days of work.
Move 1 of 5
Priced to make
34.3%
- 01
Pull your last ten finished jobs
Do this
Four numbers each. What you quoted, what it cost you, the margin you planned, the margin you got. Ten rows in a spreadsheet.
Worked example & time by hand
On the worked example
Clerkenwell was quoted at £100,270, priced to make 34.3%. It delivered 28%.
By hand
About a day. Most of that is finding the numbers, which is itself the finding.
- 02
Split every gap by cause
Do this
Not that you went over. Where. Labour, materials, subcontract, and anything a change added. One line each.
Worked example & time by hand
On the worked example
812 hours against 748 forecast. That reads as a 8.6% labour overrun, and it is the wrong answer.
By hand
Two or three hours a job.
- 03
Count what each change bought, hours as well as money
Do this
A change adds revenue. It also adds work. Put both on the job. Almost everyone logs the money and stops there, and the forecast quietly stays where it was.
Worked example & time by hand
On the worked example
The Clerkenwell change was raised, priced, approved and billed: £8,400 onto the contract. The 85 hours it bought never reached the forecast. Measured against 833, labour came in 2.5% under.
By hand
An hour, and an uncomfortable one.
- 04
Work out a correction per category, with a sign on it
Do this
One percentage per cost category, from your own ten jobs. Positive where you underprice. Negative where you overprice. This is the number a gut feel cannot give you.
Worked example & time by hand
On the worked example
Materials and site costs +23.9%. Labour -2.2%. Subcontract +1.5%. A flat +8.3% across everything would have been wrong three times over.
By hand
An hour, once the first three moves are done.
- 05
Price the next job with it, and write down what you assumed
Do this
Apply the correction where it belongs. Then record what you priced against, so the next post-mortem has something to check rather than a memory to argue with.
Worked example & time by hand
On the worked example
The next job of this shape prices materials up and labour slightly down. Neither is a guess any more.
By hand
Ongoing. It is the move everyone skips, and skipping it is why move one takes a day.
Want the worksheet?
The five moves with columns for final revenue, expected and actual costs by category, approved changes, and the correction each category needs. The totals work themselves out at the bottom.
AFTERWARDS
This is the loop we close for you
Everything above is doable by hand. Once. Doing it after every finished job is the part nobody sustains, and that is the whole of what Korrel is for. The correction updates itself as work completes, and the next estimate starts from it instead of from optimism.
The full walkthrough, from brief to sign-off to what the finished job teaches.